(CANCER RESEARCH 51, 1300-1307, February 15, 1991|

Expression of the Natural Killer Cell-associated Antigens CD56 and CD57 inHuman Neural and Striated Muscle Cells and in Their Tumors1

Gunhild Mechtersheimer, Martina Staudter, and Peter Möller2

Institute of Pathology; University of Heidelberg, Heidelberg, Federal Republic of Germany

ABSTRACT

The expression of the natural killer cell-associated antigens CD56 and

(1)57 which are known to show sequence homologies to the neural celladhesion molecule N-CAM was examined immunohistochemically in

normal and regenerative human neural and striated muscle cells and intheir tumors using monoclonal antibodies Leu-19 and Leu-7. The pattern

of expression of CD56 and CD57 in neural tissue was assessed incomparison with that of M, 68,000 neurofilament. Dendritic interstitialcells were discriminated from neural cells by application of the panleukocyte antigen CD53.

In normal tissue, CD56 was expressed in thin nerve fibers, fine varicoseand sensory nerve endings, cell membranes of ganglion cells, and fetalstriated muscle cells. Thick nerve fibers, perikarya of ganglion cells, andadult striated muscle fibers were CD56 negative. In the normal state,CD57 was restricted to thick nerve fibers. Enhanced expression orreexpression of CD56 was found in regenerative neural cells which, inpart, were also CD57 positive and in regenerative, CD57-negative striated

muscle cells. In neural tumors, CD56 was detectable in 3 of 3 benign and8 of 13 malignant schwannomas, 1 of 4 peripheral neuroepitheliomas, 4of 4 ganglioneuromas, and 8 of 8 (ganglio-)neuroblastomas, whereas

CD57 was restricted to one benign and one malignant schwannoma.Furthermore, CD56, but not CD57, could be found in 8 of 8 rhabdomyo-

sarcomas. All in all, the pattern of expression of CD56 is comparable tothat of N-CAM. CD57 exhibits a very restricted binding pattern which,

in most instances, is complementary to that of CD56. The absence ofCD56 in some neural tumors might reflect alterations in cell-cell inter

actions. Its reexpression in regenerative striated muscle cells and inrhabdomyosarcomas suggests a role of CD56 as an oncodevelopmentalantigen.

INTRODUCTION

CD56 and CD57 are newly defined clusters of leukocytedifferentiation (CD)3 antigens corresponding to the NK3 cell-associated molecules NKH-1 and HNK-1, respectively.CD57(HNK-1) comprises three antibodies one of which is Leu-7, raised against the human T-cell line HSB-2 (1). In hemato-poietic tissue, CD57 is present in 15-20% of peripheral bloodmononuclear cells, including 60% of NK-active cells, and asubset of peripheral blood cells without NK activity (2, 3).CD57 detects an A/r 110,000 glycoprotein. However, since theantibody-binding site is a carbohydrate structure, there mightbe more than one molecular structure accounting for apparentdifferences in published molecular weights. Furthermore, CD57has been found to recognize the myelin-associated antigen (4)and it also detects a sulfated carbohydrate epitope expressed ona portion of N-CAM isoforms (5).

Received 8/3/90; accepted 11/20/90.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This study was supported by a grant from the Deutsche Krebshilfe/Dr.Mildred Scheel-Stiftung fürKrebsforschung (W50/89/MÖ 2).

2To whom requests for reprints should be addressed, at Pathologisches Institutder Universität Heidelberg, Im Neuenheimer Feld 220, D-6900 Heidelberg,Federal Republic of Germany.

'The abbreviations used are: CD. cluster of designation; NK, natural killer;mAb, monoclonal antibody; AEC, 3-amino-9-ethylcarbazole: NF. neurofilament:D1C, dendritic interstitial cells; RMS, rhabdomyosarcoma.

CD56 consists of six antibodies including Leu-19. The Leu-19 antigen is an M, 220,000/140,000 cell surface glycoprotein(6-8) corresponding to the NKH-1 molecule (9, 10). It isexpressed on virtually all human natural killer cells and on asubset of T-lymphocytes and interleukin-2-activated thymo-cytes that mediate major histocompatibility complex-unrestricted cytotoxicity (6, 9, 10). Additionally, it has been foundon certain CD4+ thymocyte cell clones, some myeloid leuke-

mias, and the KG la immature hematopoietic leukemia cell line(7). Recently, Lanier et al. (8) found that CD56 was expressedalso in neural tissue and they were able to demonstrate itsidentity to the A/r 140,000 isoform of the neural cell adhesionmolecule N-CAM.

N-CAM, an integral membrane glycoprotein, is one of thebest studied cell adhesion molecules (11). It is a member of theimmunoglobulin supergene family (12) and mediates Ca2+-

independent homophilic intercellular binding (11). A considerable number of molecular isoforms have been found whichmake up two main categories, namely, brain and muscle N-CAM (12-14). Furthermore, soluble forms of N-CAM havebeen isolated (15, 16). It is now known that these major isoforms are generated from a single-copy gene located on bandq23 of human chromosome 11 (17) by alternative RNA processing at the splicing and polyadenylation levels (12, 18). Theydiffer mainly in the amounts of sialic acid at the carboxy-terminal ends of the molecules, reflecting different stages duringembryogenesis (19).

In the normal state, N-CAM is present in all three germlayers during early development (20). In the adult, however, itis mainly restricted to neural cells (11). Thus, N-CAM has beenfound to play a central role in various developmental eventsincluding the orderly outgrowth of axons (21), cell patternformation (19), and nerve-muscle interactions (22). Furthermore, it is assumed that cell-cell interactions mediated by N-CAM act as regenerative key mechanisms since N-CAM ispresent in regenerative neural and striated muscle tissues (23,24).

Although the distribution pattern of N-CAM has been extensively studied in vertebrates, information concerning its patternof expression in human tissue is still scarce. The identity ofCD56 and CD57 molecules with the M, 140,000 isoform andwith at least a portion of N-CAM, respectively, prompted us tostudy their expression in a comprehensive series of normalneural and striated muscle cells and in related tumors by application of mAbs Leu-19 and Leu-7, respectively.

MATERIALS AND METHODS

Tissue. Most tissues, including all tumors, were obtained from surgical specimens. Some nonneoplastic tissues were drawn from autopsymaterial during the first 6 h after death. Diagnosis of the tumors wasbased on standard histopathological criteria as described by Enzingerand Weiss (25). The tissue samples were frozen quickly in liquidnitrogen and stored at —¿�70°C.Serial frozen sections of about 1 cm2

and a thickness of 4-6 urn were dried in air, fixed in acetone at roomtemperature for 10 min, and immediately stained or stored at —¿�20°C

for a short period.

1300

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

Reagents. mAb Leu-19 (isotype IgGl) recognizing the CD56(NKH-1) antigen and mAb Leu-7 (isotype IgM) directed against theCD57(HNK-1) antigen were obtained from Becton-Dickinson (Mountain View, CA); mAb NR4 (isotype IgGl) directed against M, 68,000neurofilament was purchased from Dakopatts (Copenhagen, Denmark).MAb HD77 [isotype IgM(Gl)] recognizing a recently defined panleukocyte antigen designated CD53 (26) was raised in our own laboratory (in collaboration with Dr. B. Dörkenand Dr. G. Moldenhauer).CD22 mAb HD39 (isotype IgGl) and CD21 mAb B2 (isotype IgM).which were used as isotype-matched control antibodies, were providedby B. Dörkenand G. Moldenhauer (Heidelberg, Federal Republic ofGermany) and by L. N. Nadler (Boston, MA), respectively. A polyclonalbiotinylated sheep antibody to mouse immunoglobulin (reactive withall mouse isotypes) and a streptavidin-biotinylated peroxidase complexwere provided by Amersham (High Wycombe, United Kingdom). AECand N'N-dimethylformamide were obtained from Sigma Chemical Co.

(St. Louis, MO).Staining Procedure. After rehydration with phosphate-buffered saline

solution (pH 7.5), the frozen sections were incubated for l h withpurified mAbs at appropriate dilutions (Leu-19, 1:20; Leu-7, 1:20;NR4, 1:50; HD77, 1:100). The sections were then incubated withbiotinylated anti-mouse immunoglobulin (1:50) and streptavidin-biotinylated peroxidase complex (1:100) for 30 min, respectively. All incubation steps were carried out in a humid chamber at room temperatureand then rinsed twice in phosphate-buffered saline. The peroxidasereaction product was reddish brown when AEC and H2Oz were used assubstrate: 4 mg AEC was dissolved in 500 n\ yVW-dimethylformamide.Subsequently, 9.5 ml sodium acetate buffer (0.05 M; pH 5.0) and, justbefore use, 5 n\ H2Û2(30%) were added. The incubation time was 15min. The sections were then rinsed in distilled water, counterstainedwith Harris' hematoxylin, and mounted in glycerol gelatin.

Controls. Negative controls in each case were performed without theprimary antibody; no staining was observed except for scattered gran-ulocytes. This staining was due to endogenous peroxidase, which wasnot blocked for the benefit of optimal antigenicity. The positive resultswere confirmed by controls using the isotype-matched mAbs HD39(1:100) and B2 (1:100), which yielded negative results in all cases,ruling out nonspecific antibody binding. Minor peripheral nerve trunkspresent in almost all tissue samples studied served as positive intrinsiccontrols for the reactivity of CD56, CD57, and NF mAbs, respectively.Furthermore, expression of CD56 and, to a lesser extend, of CD57 inscattered CD53-positive lymphoid cells proved the immune reaction tobe reliable.

Evaluation. Detailed expression of CD56 and CD57 antigens inperipheral nerves, sensory nerve endings, autonomie ganglia, and neuraltumors was assessed in comparison with the reactivity of NF in thesetissues. mAb CD53(HD77) was used, and the visualization of DIC, inaddition to lymphoid cells, which were abundant especially in neuraltissue, allowed a clear discrimination between nonneoplastic and neo-plastic neural cells. In neoplasias, the staining of tumor cells wasevaluated in a semiquantitative fashion: +, entire tumor cell populationpositive; ±,positive and negative tumor cells in various amounts (fordetails see "Results"); -, entire tumor cell population negative.

RESULTS

The expression of CD56 and CD57 antigens in nonneoplasticneural and striated muscle cells as well as in their tumors isshown in Tables 1 and 2.

Normal Neural Tissues. The analysis of normal peripheralnerve trunks showed expression of CD56 in thin CD57~ nervefibers. Thick CD57+ nerve fibers lacked any detectable CD56

molecules (Fig. 1, A and B). Both thin and thick nerve fiberswere NF+ (Fig. 1C). Staining with mAb CD53 showed a con

siderable number of DIC within the major nerve trunks andthus allowed clear discrimination between neural cells and DIC(Fig. ID). Minor peripheral nerve trunks present in varioustissues exhibited the same antigenic profile as major peripheralnerve trunks. In addition, fine varicose nerve fibers surrounding

Table 1 Expression ofCD56 and CD57 antigens in nonneoplastic neural andstriated muscle cells

Cell type CD56 CD57Thin peripheral nerve fibers +"

Thick peripheral nerve fibers +Fine varicose nerve fibers +Regenerative nerve fibers + +/-Meissner's corpuscles +

Pacinian corpuscles (axons) +Pacinian corpuscles (lamellae)Intrafusal muscle fibers +/—Ganglion cells + —¿�Satellite cells +Chromaffine cells (adrenal medulla) + +/(+)Striated muscle fibers (first trimenon) +Striated muscle fibers (third trimenon) (+)/-Striated muscle fibers (adult) -/(+)*

Striated muscle fibers (regenerative) +/(+)"Scoring of cell reaction: +, strong staining intensity for all cells; (+), weak

staining intensity for all cells; •¿�. positive and negative cells in various amounts;-, negativity for all cells.

* Weak positivity in some muscle fibers of eye muscles and of the tongue.

Table 2 Expression ofCD56 and CD57 in tumors of neural and striatedmuscle origin

CD56/CD57

PhenotypeBenign

schwannomaMalignantschwannomaPeripheralneuroepitheliomaGanglioneuromaGanglion

cellsSchwanncellsGanglioneuroblastomaGanglion

cellsNeuroblastsNeuroblastomaRhabdomyosarcomaN

+°3

3/-135/-4_/_44/—4/-4-/44/—4

4/-86/-+/--/I3/1l/--/-—

/34/——

/——/-2/--~/25/123/4—

/4-/I—

/4-/4—

/4-/8

°Scoring of tumor cell reaction: +, strong positivity for all tumor cells; +/-strongly and/or weakly positive and negative tumor cells in various amounts; -negativity for all tumor cells.

cutaneous appendages and the base of crypt cells in a boweltissue specimen expressed CD56 antigens in the absence ofCD57 and NF molecules (Fig. JA). Meissner's corpuscles andaxons of Pacinian corpuscles were CD56+ but CD57~ (Fig. 2,

B and C); they also expressed NF. Surrounding lamellae composed of flattened spindle cells were CD56~/CD57~/NF~. The

perikarya of ganglion cells of peripheral ganglia were stronglyNF+ but lacked any detectable CD56 and CD57 molecules; thesatellite cells were CD56+ but CD57~ and NF~ (Fig. 3). Scat

tered DIC surrounding the ganglia were discriminated fromCD56* satellite cells by expression of CD53 antigen. Expres

sion of CD56 on the surface of ganglion cells was difficult toassess because of the strong CD56 positivity of surroundingsatellite cells. Yet, in ganglioneuromas and optionally also inganglioneuroblastomas, immature ganglion cells lacking satellite cells clearly showed membrane-bound staining for CD56molecules (cf. below). Chromaffine cells of the adrenal medullawere CD56+CD57+/(+)/NF+. Altogether, CD56 and CD57 dis

played in most instances a complementary pattern of expressionwith an overall predominance of CD56.

Regenerative Neural Cells. In traumatic neuromas the overwhelming majority of haphazardly oriented nerve fibers wereCD56* (Fig. 4A). In addition, and in contrast to the normal

state, serial section preparations disclosed expression of CD56antigen also in at least some of the CD57* nerve fibers (Fig.

4B). NF was consistently expressed in traumatic neuromas.Tumors of the Peripheral Nervous System. In three benign

schwannomas, the entire spindle tumor cell population wasCD56+. CD57 and NF, however, were restricted to a tumor cell

1301

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

ÕA M

. 'p '¡.a*y''•y

'.¿A1BI ,t »•

' //•*'•/"'j/f,lfc'//' r ' ,! <Wlfffit ;/>, ''

"-W' ' i •¿�<' ¿/ '"' '<!wt,A '••/¿' (f¿' •¿�/ ^Â¥'J*> ¿>^ V '•-¿,'4jw* •¿�->,/ <" **Affa . >-/\ Ci* '

'.*

.'' ">

t

*>-S*

Fig. 1. Serial section preparation of a major peripheral nene trunk (peronealnene). In A. CD56 detected by mAb Leu-19 is confined to thin nerve fibers. InB, mAb Leu-7, in contrast, shows expression of CD57 in the myelin sheath ofthick nene fibers. In C, a neurofilament visualized by mAb NR4 is expressed inboth axons of thick and thin nerve fibers. In D, CD53 illustrated by mAb HD77allows a clear discrimination between nene fibers and dendritic interstitial cells.Immunoperoxidase staining of frozen sections, faint hematoxylin counterstain;same technique for all photographs; x 145.

subpopulation of one case, respectively. Five of 13 malignantschwannomas exhibited strong and consistent expression ofCD56 in the entire tumor cell population (Fig. 5/1); in 3 of 13tumors at least one-half of the neoplastic cells were CD56+

(Figs. 5B and 9Ä),whereas 5 of 13 cases lacked any detectableCD56 molecules. CD57 and NF were each restricted to onecase within a minor tumor cell subpopulation. Three peripheralneuroepitheliomas were CD56~ throughout (Fig. 8/1); in one

case some scattered undifferentiated tumor cells expressedCD56 antigen. CD57, like NF, was absent in all peripheralneuroepitheliomas studied. Thin CD56+ and thick CD57* nervefibers, which were NF+, and/or scattered CD56+/CD57+

lymphoid cells served as positive intrinsic controls for theimmune reaction in CD56" and/or CD57~ tumors, ruling out

false-negative staining (Fig. 5C).Tumors of Autonomie Ganglia. Some of the ganglion cells of

four ganglioneuromas studied lacked satellite cells and showedclearly discernible CD56+ cell membranes; mature ganglioncells in addition showed CD56+/CD57~/NF~ satellite cells. Themajority of Schwann cells were CD56+, whereas CD57 wasonly present in about one-half of the Schwann cell compartmentin one ganglioneuroma and in two cases within a minorSchwann cell population; at least some of the CD57+ Schwann

cells coexpressed CD56. NF was consistently expressed in theSchwann cells of ganglioneuromas. Immature ganglion cellspresent in four ganglioneuroblastomas lacked satellite cells andshowed CD56'/CD57~/NF+ perikarya; some but not all NF+

ganglion cells exhibited membrane-bound staining for CD56,as did all CD57~/NF" small, round, undifferentiated neuro-

blasts in these cases as well as in four undifferentiated neuro-blastomas (Fig. 6). Additionally, extracellular CD56+ deposits

were found around some larger neuroblasts (Fig. 6/1).Normal Striated Muscle Cells. Fetal skeletal muscle cells of

18 weeks of gestation were strongly CD56+ and those of 35weeks were weakly CD56+ compared to strongly CD56+ thin

nerve fibers of adjacent minor peripheral nerve trunks (Fig.1A). Adult skeletal muscle cells were CD56~ and CD57~

throughout (Fig. IB and C). A minority of muscle fibers of thetongue and of eye muscles which were extremely rich in minorperipheral nerve bundles and fine varicose nerve fibers wereCD56+ in the absence of CD57. In muscle spindles included inan upper extremity skeletal muscle, at least some of the intra-fusal muscle fibers were CD56+ and lacked CD57 molecules

(Fig. IB and C).

2B 2C

Fig. 2. Skin specimen of the finger tip. CD56 is strongly expressed in A, a dense network of fine varicose nerve fibers surrounding cutaneous appendages and bloodvessels (x 145). B, Meissner's corpuscles found in dermal papillae (x 145): C, axons of Pacinian corpuscles while absent in surrounding flattened lamellae (x 90).

1302

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

* .

seFig. 3. Serial section preparation of a sympathetic ganglion (thorax). In A, the perikarya of ganglion cells and adjacent nerve fibers are strongly neurofilament-

positive, whereas satellite cells lack any detectable neurofilament. In B, in contrast, CD56 is expressed on the cell surface and on satellite cells of ganglion cells. In C,CD57 is confined to myelin sheaths of thick nerve fibers present in an adjacent minor peripheral nene trunk (arroH's) and to scattered lymphoid cells, x 181.

Damaged Striated Muscle Cells. The great majority of striatedmuscle fibers microtopographically associated with malignanttumor cells were CD56+ in all seven cases studied (Fig. &4), as

were most of the muscle fibers associated with benign tumor

Fig. 4. Traumatic neuroma (ischiadic nerve). In A. nearly all hapha/ardlyoriented nerve fibers express CD56 molecules. In B, in contrast to the normalstate, some of the CD56-positive nerve fibers also show CD57-positive myelinsheaths. X 145.

cells in one of two cases. In one skeletal muscle which showedextensive fibrosis due to previous operative intervention, mostof the muscle cells were CD56+ (Fig. 8Ä),as were muscle fibers

of a skeletal muscle biopsy specimen which showed necrosiscaused by vascular embolization. Like in normal striated musclecells, CD57 was absent in all damaged striated muscle cells.

Tumors of Striated Muscle Origin. Included were six embryonal, one alveolar, and one pleomorphic RMS. In fourembryonal, in the alveolar, and in the pleomorphic RMS, alltumor cells, i.e., small, round, undifferentiated tumor cells aswell as rounded or elongated rhabdomyoblasts present in various amounts, were strongly and consistently CD56+ (Fig. 9A).

In one embryonal RMS a major tumor cell population expressed CD56 molecules; in another case, which was exclusivelycomposed of rounded rhabdomyoblasts, a minor tumor cellpopulation also expressed CD56 molecules. All RMS studiedwere CD57~.

DISCUSSION

There is increasing evidence for mAbs to detect identicalepitopes that might be part of different structures. This holdstrue for some of the CD antigens which may be found in variousnormal and/or neoplastic extrahematopoietic cells, either con-stitutively or resulting from malignant transformation.CD57(HNK-1) is one of the best studied antigens. Outside thehematopoietic system it has been found in a series of neural orneuroectodermal tumors, including neurofibromas and neuro-fibrosarcomas (27, 28), malignant melanomas (29, 30), malignant peripheral neuroectodermal tumors (30, 31 ), Ewing's sar

comas (32), and even small cell lung carcinomas and otherneuroendocrine tumors (33). This seems to reflect the fact thatCD57 also recognizes myelin-associated glycoprotein and, additionally, a sulfated carbohydrate epitope expressed on a portion of N-CAM isoforms (5).

CD56(NKH-1), which belongs also to the NK-associated

antigens, has recently been demonstrated to be identical to theM, 140,000 isoform of N-CAM (8). This prompted us to study

the expression of CD56 and CD57 in human neural and striatedmuscle cells as well as in their tumors by application of mAbsLeu-19(8)andLeu-7(l).

Our results on the expression of CD56 and CD57 antigensin normal human nerve fibers are in parallel with data reportedby Nieke and Schachner (34) who, in adult rats, found N-CAM

1303

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 [N HUMAN NEURAL AND STRIATED MUSCLE TISSUES

Fig. 5. A, spindle cell malignant schwannoma of the buttock. CD56 is detectable in the entire tumor cell population. B, pleomorphic malignant schwannoma ofthe retroperitoneum. About one-half of the tumor cell population lacks any detectable CD56 molecules. C, spindle cell malignant schwannoma of the stomach. Thetumor cells are CD57-negative throughout. Scattered CD57-positive lymphoid cells serve as an intrinsic control for the immune reaction, x 181.

Fig. 6. /i, ganglioneuroblastoma of the adrenal gland. Small, round, undiffer-entiated neuroblasts, scattered immature ganglion cells (arrows), and neuropil-like processes show strong expression of CD56 antigens. Additionally, extracellular deposits of CDS6 are found around larger neuroblasts (stars), x 145. B,neuroblastoma of the thoracic wall. The entire undifferentiated tumor cell population is strongly CDSo-positive, while endothelial cells are unstained, x 181.

immunoreactivity predominantly associated with unmyelinatedSchwann cells, whereas the CD57(HNK-1) epitope appeared tobe associated with the outer profile of thick myelin sheaths, afinding which was also been described by Smolle et al. (27) whoinvestigated the pattern of CD57 expression in a variety of skintumors. In traumatic neuromas and in the Schwann cell compartment of ganglioneuromas, however, CD56 was morebroadly distributed and could also be found in some CD57-positive Schwann cells. In the distal end of transected ratischiadic nerves, N-CAM also reappeared in many Schwanncells (34-36); however, N-CAM was restricted to CD57-nega-tive nerve fibers. Like in traumatic neuromas and in theSchwann cell compartment of ganglioneuromas, CD56 wasconsistently expressed in benign schwannomas whereas, in contrast to data, CD57 and/or neurofilament were only rarelydetectable in these neoplasms and in their malignant counterparts (27, 28). Malignant schwannomas, however, exhibited aheterogenous pattern of CD56 expression, i.e., some cases werepartially or even completely CD56-negative, reflecting the absence of at least the M, 140,000 isoform of N-CAM in thesetissues. Since N-CAM is required for normal adhesive interactions between cells, low amounts or even absence of N-CAMmight lead to a loss of contact inhibition allowing cell detachment and migration in an uncontrolled manner. In line withthis hypothesis, Linnemann et al. (37) recently demonstratedthat the metastatic melanoma cell line K1735-M1 expressedless N-CAM than did the nonmetastasizing K1735-C116 line.Furthermore, mAb MUC18, a marker for tumor progressionin human melanoma, has been found to show sequence similarity to neural cell adhesion molecules including N-CAM (38).Additionally, N-CAM expression is markedly reduced aftertransformation of retinal and neural cells with Rous sarcomavirus (39, 40).

N-CAM expression in neuroblastomas was initially describedin mice showing a shift to the predominant expression of theA/r 180,000 isoform in more differentiated neuroblasts (41).Since the M, 180,000 isoform of N-CAM occurs at later stagesduring development than do the M, 120,000 and 140,000isoforms, it is suggested that A/r 180,000 N-CAM plays adifferential role in the stabilization of cell contacts. N-CAMexpression in human neuroblastoma cell lines was first de-

1304

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

:7C •¿� v .-

Fig. 7. In A, a fetal skeletal muscle of 18 weeks of gestation exhibits strong surface expression of CD56 in all muscle fibers. B and C, serial section preparation ofan adult skeletal muscle. In B, the muscle fibers lack any detectable CD56. Intrafusal muscle fibers of a muscle spindle are at least partially CDSo-positive, as aresome small nerve fibers. In C, CD57 is confined to myelin sheaths of CDSo-negative nerve fibers, x 181.

scribed by Lipinski et al. (42) using a rabbit anti-humanN-CAM antiserum. We found expression of CD56 in all (gang-lio)neuroblastomas studied. Feickert et al. (43), using the sameantibody we applied, found in situ expression of M, 140,000 N-

8B

Fig. 8. A, peripheral neuroepithelioma of the thoracic wall. The undifferen-tiated tumor cells are consistently CD56-negative, whereas infiltrated skeletalmuscle fibers are strongly CD56-positive. x 181. A, skeletal muscle of the lowerextremity showing extensive fibrosis due to previous operative intervention. Thegreat majority of muscle fibers are strongly CD56-positive. x 145.

CAM in only 6 of 11 neuroblastomas, whereas CD56 mAb T-199 yielded positive results in all cases studied. Since CD56mAb only detects the M, 140,000 isoform of N-CAM, theCD56 negativity of some neuroblastomas might reflect differences in the maturation grade of the various neoplasms studied.However, our cases, which included different grades of maturation, were all CD56 positive. In contrast, CD56 was onlyrarely detectable in peripheral neuroepitheliomas. Since N-CAM has been found in a series of Ewing's sarcomas both insitu and in vitro (42, 44, 45) and since Ewing's sarcoma and

peripheral neuroepithelioma are considered to be related neoplasms based on the shared reciprocal t(ll;22) translocation(46) and the shared protooncogene expression (47), the rareCD56 expression in peripheral neuroepitheliomas is noteworthy. Another remarkable feature was the consistent absence ofCD57 in all (ganglio-)neuroblastomas studied. This contrastswith data by Caillaud et al. (30) who found CD57 expressionin 11 of 13 neuroblastomas studied. However, since in ourhands some CD57-positive lymphoid cells and/or thick nervefibers showed reliable immune reaction and since mAb Leu-7is of IgM isotype, the considerable number of CD57-positiveneuroblastomas reported in the literature might at least partiallybe the result of unspecific isotype effects which, e.g., can befound in several epithelial cell types (data not shown).4

The knowledge of the distribution pattern of N-CAM inhuman tissue is further enriched by results obtained with somemAbs of the neuroblastoma and the small cell lung cancerworkshops, which have been found to also share an epitope incommon with N-CAM. Thus, mAbs UJ13A and 3F8 have beendemonstrated in situ in 4 of 4 neuroblastomas and in vitro in 6of 8 neuroblastoma cell lines, respectively (45, 48). In additionto sharing an epitope in common with N-CAM, mAb 3F8 alsorecognizes the ganglioside GD2 (49, 50) which appears to be aparticularly good marker for neuroectodermal tumors such asneuroblastoma and malignant melanoma. Both UJ13A and 3F8have gained diagnostic and/or therapeutic application (51-53).Given that N-CAM is present in normal peripheral nerves andautonomie ganglia, the diagnostic and therapeutic value ofmAbs recognizing an epitope in common with N-CAM shouldbe reassessed. In addition to their expression in neuroblastomasand malignant melanomas, some of the mAbs sharing an epitope in common with N-CAM including CD56 mAb have alsobeen demonstrated in medulloblastomas, Ewing's sarcomas,

4 P. Möllerand K. Koretz, unpublished observations.

1305

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

. . r•¿�rf&feö . ««••.1*;.'->

Fig. 9. A, embryonal rhabdomyosarcoma of the retroperitoneum. Small, round,undifferentiated tumor cells as well as rounded or elongated rhabdomyoblastsdisplay strong staining for CD56. Myxoid matrix is unstained, x 145. B, malignant schwannoma with rhabdomyoblastic differentiation (Triton tumor) of thelower extremity. The majority of tumor cells, including some with rhabdomyoblastic differentiation (arrows), are CD56 positive. xl81.

Wilms' tumors, and even in rhabdomyosarcomas (42-45, 48).

In line with these results, Roth et al. (54) found evidence of thelong chain form of polysialic acid of N-CAM in Wilms' tumors.

Since N-CAM is present in fetal but absent in adult kidneys,

the authors suggested that this molecule might represent anoncodevelopmental antigen in renal tissue. Like mAbs of theneuroblastoma and small cell lung cancer workshops, CD56was expressed in rhabdomyosarcomas, whereas CD57 was completely absent in all tumors of striated muscle origin. In contrastto the reports of Feickert et al. (43), who showed in situexpression of CD56 in 2 of 6 rhabdomyosarcomas only, wefound CD56 expression in variable amounts in 8 of 8 rhabdomyosarcomas studied. Because N-CAM is present in fetal butabsent in adult striated muscle cells in the normal state andreexpressed in rhabdomyosarcomas, we suggest that, in striatedmuscle tissue, this antigen might also be considered an oncodevelopmental antigen. Reexpression of N-CAM in regenerativemuscle fibers in various myopathies (55, 56) is in line with thisconclusion because regenerative muscle fibers also follow thedevelopmental pathway. Additionally, we found reexpressionof CD56 in regenerative muscle fibers in damaged skeletalmuscles and in skeletal muscle fibers microtopographically

associated with tumor cells. All in all, the distribution patternof CD56 in neural and striated muscle cells and, according toprevious and recent literature, in their tumors corresponds tothat of N-CAM. CD57, in contrast, shows a pattern of expression which in most instances is complementary to that of CD56and thus far from covering that of N-CAM. The absence ofCD56 in some malignant schwannomas and peripheral neuroe-pitheliomas might reflect alterations in cell-cell interactions.CD56 expression in normal fetal and regenerative striatedmuscles, as well as in rhabdomyosarcomas, suggests a role ofthis molecule as an oncodevelopmental antigen, since it isabsent in normal adult striated muscles. These results obtainedin striated muscle cells implicate parallels to those found inrenal tissue. Since N-CAM is expressed in normal peripheralnerves and autonomie ganglia, the diagnostic and therapeuticvalue of mAbs sharing an epitope in common with N-CAM hasto be reassessed.

ACKNOWLEDGMENTS

The skillful technical assistance of Inge Brandt and Sybille Mengesand the excellent photographic assistance of John Moyers are gratefullyacknowledged. We are indebted to Christine Raulfs for editing themanuscript.

REFERENCES

1. Abo, T., and Balch, C. M. A differentiation antigen of human NK and Kcells identified by a monoclonal antibody (HNK-1). J. Immunol., 127: 1024-1029, 1981.

2. Lanier, L. L., Le, A. M., Phillips, J. H., Warner, N. L., and Babcock, G. F.Subpopulations of human natural killer cells defined by expression of theLeu-7 (HNK-1) and Leu-11 (NK-15) antigens. J. Immunol., 131:1789-1796,1983.

3. Manara, D. C., de Panfilis, G., and Ferrari, C. Ultrastructural characterization of human large granular lymphocyte subsets defined by the expressionof HNK-1 (Leu-7), Leu-11, or both HNK-1 and Leu-11 antigens. J. Histo-chem. Cytochem., 33: 1129-1133, 1985.

4. McGarry, R. C., Heifand, S. L., Quarles. R. H., and Roder, J. C. Recognitionof myelin-associated glycoprotein by the monoclonal antibody HNK-1. Nature (Lond.), 306: 376-378, 1983.

5. Kruse, J., Mailhammer, R., Wennecke, H., Faissner, A.. Sommer, I., Goridis,C., and Schachner, M. Neural cell adhesion molecules and MAG share acommon carbohydrate moiety recognized by monoclonal antibodies L2 andHNK-1. Nature (Lond.), 311: 153-155, 1984.

6. Lanier, L. L., Le, A. M., Civin, C. I., Loken, M. R., and Phillips, J. H. Therelationship of CD16 (Leu-11) and Leu-19(NKH-l) antigen expression onhuman peripheral blood NK cells and cytotoxic T lymphocytes. J. Immunol.,136: 4480-4486, 1986.

7. Lanier, L. L., Le, A. M., Ding, A., Evans, E. L., Krensky, A. M., Clayberger,C., and Phillips, J. H. Expression of Leu-19 (NKH-1) antigen on human II-2-dependent cytotoxic and non-cytotoxic T cell lines. J. Immunol., 138:2019-2023, 1987.

8. Lanier, L. L., Testi, R., Bindl, J., and Phillips, J.H. Identity of Leu-19(CD56)leucocyte differentiation antigen and neural cell adhesion molecule. J. Exp.Med., 769:2233-2238, 1989.

9. Griffin, J. D., Hercend, T., Beveridge, R. P., and Schlossman, S. F. Characterization of an antigen expressed by human natural killer cells. J. Immunol.,130: 2947-2951, 1983.

10. Hercend, T., Griffin, J. D., Bensussan, A., Schmidt, R. E., Edson, M. A.,Brennan, A., Murray, C, Daley, J. F., Schlossman, S. F., and Ritz, J.Generation of monoclonal antibodies to a human natural killer clone. Characterization of two natural killer-associated antigens, NKH1A and NKH2,expressed on subsets of large granular lymphocytes. J. Clin. Invest., 75:932-943, 1985.

11. Edelman, G. M. Cell adhesion molecules in the regulation of animal formand tissue pattern. Annu. Rev. Cell Biol., 2: 81-116, 1986.

12. Cunningham, B. A., Hemperly, J. J., Murray, B.. Prediger, E. A., Bracken-bury, R., and Edelman, G. M. Neural cell adhesion molecule: structure,immunoglobulin-like domains, cell surface modulation, and alternative RNAsplicing. Science (Washington DC), 236: 799-806, 1987.

13. Dickinson, G., Gower, H. J., Barton, H., Prentice, M., Elsom, V. L., Moore,S. E., Cox, R. D., Quinn, C., Putt, W., and Walsh, F. S. Human muscleneural cell adhesion molecule (N-CAM): identification of a muscle-specificsequence in the extracellular domain. Cell, 50: 1119-1130, 1987.

14. Moore, S. E., Thompson, J., Kirkness, V., Dickinson, J. G., and Walsh, F.S. Skeletal muscle neural adhesion molecule (N-CAM): changes in proteinand mRNA species during myogenesis of muscle cell lines. J. Cell Biol., 105:1377-1386, 1987.

1306

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

CD56 AND CD57 IN HUMAN NEURAL AND STRIATED MUSCLE TISSUES

15. Gower, H. J., Barton, C. H., Elsom, V. L., Thompson, J., Moore, S. E.,Dickinson, G., and Walsh, F. S. Alternative splicing generates a secretedform of N-CAM in muscle and brain. Cell. 55: 955-964, 1988.

16. Nybroe, O., Linnemann, D., and Bock, E. Heterogeneity of soluble neuralcell adhesion molecule. J. Neurochem., 53: 1372-1378, 1989.

17. N'Guyen, C., Mattei, M.-G., Mattei, J.-F., Santoni. M.-J.. Goridis. C.. andJordan, B. R. Localization of the human NCAM gene on band q23 ofchromosome 11: the third gene coding for a cell interaction molecule mappedto the distal portion of the long arm of chromosome 11. J. Cell Biol., 102:711-715, 1986.

18. Goridis. C., and Wille, W. The three size classes of mouse NCAM proteinsarise from a single gene by a combination of alternative splicing and use ofdifferent polyadenylation sites. Neurochem. Int., 12: 269-272, 1988.

19. Chuong, C.-M., and Edelman, G. M. Alterations in neural cell adhesionmolecules during development of different regions of the nervous system. J.Neurosci., 4: 2354-2368, 1984.

20. Crossin, K. L., Chuong, C.-M., and Edelman, G. M. Expression sequence ofcell adhesion molecules. Proc. Nati. Acad. Sci. USA, 82: 6942-6946. 1985.

21. Thanos, S., Bonhoeffer, F., and Rutishauser, U. Fiber-fiber interactions andtectal cues influence the development of the chicken retinotectal projections.Proc. Nati. Acad. Sci. USA, 81: 1906-1910, 1984.

22. Grumet, M., Rutishauser, F., and Rutishauser, U. Neural cell adhesionmolecule is on embryonic muscle cells and mediates adhesion to nerve cellsin vitro. Nature (Lond.), 295: 693-695. 1982.

23. Covault, J., and Sanes, J. R. Neural cell adhesion molecule (N-CAM)accumulates in denervated and paralyzed skeletal muscle. Proc. Nati. Acad.Sci. USA, 82: 4544-4548, 1985.

24. Daniloff, J. K., Levi, G., Grumet, M., Rieger, F., and Edelman, G. M. Alteredexpression of neuronal cell adhesion molecules induced by nerve injury andrepair. J. Cell Biol., 103:929-945, 1986.

25. Enzinger. F. M., and Weiss. S. W. Soft Tissue Tumors. St. Louis: CV MosbyCo., 1988.

26. Hadam, M. R. Cluster report: CD53. In: W. Knapp. B. Dörken,W. R. Gilks,E. P. Rieber, R. E. Schmidt, H. Stein, and A. E. G. K. von dem Borne (eds.),Leucocyte Typing. IV. White Cell Differentiation Antigens, pp. 674-678.Oxford, England: Oxford University Press, 1989.

27. Smolle, J., Walter, G. F., and Kerl, H. Myelin-associated glycoprotein inneurogenic tumours of the skin: an immunohistological study using Leu-7monoclonal antibody. Arch. Dermatol. Res., 277: 141-142. 1985.

28. Swanson, P. E., Manivel, J. C., and Wick, M. R. Immunoreactivity for Leu-7 in neurofibrosarcoma and other spindle cell sarcomas of soft tissue. Am.J. Pathol., 126: 546-560, 1987.

29. Lipinski, M., Braham, K., Caillaud, J.-M., Carlu, C., and Tursz, T. HNK-1antibody detects an antigen expressed on neuroectodermal cells. J. Exp.Med., 158: 1775-1780, 1983.

30. Caillaud, J.-M., Benjelloun, S.. Bosq. J.. Braham, K., and Lipinski. M. HNK-1-defined antigen detected in paraffin-embedded neuroectoderm tumors andthose derived from cells of the amine precursor uptake and decarboxylationsystem. Cancer Res., 44: 4432-4439, 1984.

31. Llombard-Bosch, A., Lacombe, M. J., Peydro-Olaya, A., Perez-Bacete, M.,and Contesso, G. Malignant peripheral neuroectodermal tumours of boneother than Askin's neoplasms: characterization of 14 new cases with immu-

nohistochemistry and electron microscopy. Virchows Arch. A Pathol. Anat.,4/2:421-430, 1988.

32. Pinto. A., Grant, L. H., Hayes, F. A., Schell, M. J., and Parham. D. M.Immunohistochemical expression of neuron-specific enolase and Leu 7 inEwing's sarcoma of bone. Cancer (Phila.). 64: 1266-1273. 1989.

33. Bunn, P. A.. Linnoila, I., Minna, J. D., Carney, D., and Gazdar, A. F. Smallcell lung cancer, endocrine cells of the fetal bronchus, and other neuroendo-crine cells express the Leu-7 antigenic determinant present on natural killercells. Blood, 65: 764-768, 1985.

34. Nieke, J., and Schachner, M. Expression of the neural cell adhesion moleculesLI and N-CAM and their common carbohydrate epitope L2/HNK-I duringdevelopment and after transection of the mouse sciatic nerve. Differentiation,30: 141-151, 1985.

35. Martini, R., and Schachner, M. Immunoelectron miscroscopic localizationof neural cell adhesion molecules (LI, N-CAM, and MAG) and their sharedcarbohydrate epitope and myelin basic protein in developing sciatic nerve. J.Cell Biol., 103: 2439-2448, 1986.

36. Martini, R., and Schachner, M. Immunoelectron microscopic localization ofneural cell adhesion molecules (LI, N-CAM, and myelin-associated glycoprotein) in regenerating adult mouse sciatic nerve. J. Cell Biol., 106: 1735-1746, 1988.

37. Linnemann, D., Raz, A., and Bock, E. Differential expression of cell adhesion

molecules in variants of K1735 melanoma cells differing in metastatic capacity. Int. J. Cancer, 43: 709-712, 1989.

38. Lehmann, J. M., Riethmuller, G., and Johnson. J. P. MUC18, a marker oftumor progression in human melanoma, shows sequence similarity to theneural cell adhesion molecules of the immunoglobulin superfamily. Proc.Nati. Acad. Sci. USA. 86: 9891-9895, 1989.

39. Brackenbury, R., Greenberg, M. E., and Edleman. G. M. Phenotypic changesand loss of N-CAM-mediated adhesion in transformed embryonic chickenretinal cells. J. Cell Biol., 99: 1944-1954, 1984.

40. Greenberg M. E., Brackenbury R., and Edelman G. M. Alteration of neuralcell adhesion molecule (N-CAM) expression after neuronal cell transformation by Rous sarcoma virus. Proc. Nati. Acad. Sci. USA, 81: 969-973, 1984.

41. Pollerberg, E. D., Sadoul, R., Goridis, C., and Schachner, M. Selectiveexpression of the 180-kDa component of the neural cell adhesion moleculeN-CAM during development. J. Cell Biol.. 101: 1921-1929, 1985.

42. Lipinski, M., Hirsch, M.-R., Deagistini-Bazin, H., Yamada, O.. Tursz, T.,and Goridis, C. Characterization of neural cell adhesion molecules (NCAM)expressed by Ewing and neuroblastoma cell lines. Int. J. Cancer, 40: 81-86,1987.

43. Feickert, H.-J., Pietsch, T., Hadam, M. R., and Riehm, H. NK-cell marker.mAb T-199 detects a new antigenic determinant distinct from the N901,Leu-19, and Leu-7 antigens or antigen epitopes expressed on NK-cells. In:W. Knapp, B. Dörken,W. R. Gilks, E. P. Rieber, R. E. Schmidt, H. Stein,and A. E. G. K. von dem Borne (eds.). Leucocyte Typing. IV. White CellDifferentiation Antigens, pp. 705-708. Oxford, England: Oxford UniversityPress, 1989.

44. Patel, K., Roussel, R. J., Pemberton, L. F., Cheung, N. K., Walsh, F. S.,Moore, S. E., Sugimoto, T., and Kemshead, J. T. Monoclonal antibody 3F8recognized the neural cell adhesion molecule (NCAM) in addition to theganglioside GD2. Br. J. Cancer. 60: 861-866, 1989.

45. Patel, K. Roussel, R. J., Bourne, S., Moore, S. E., Walsh, F. S., andKemshead, J. T. Monoclonal antibody UJ13A recognizes the neural celladhesion molecule (NCAM). Int. J. Cancer. 44: 1062-1068, 1989.

46. Whang-Peng, J., Triche, T. J., Knutsen, T., Miser, J.. Kao-Shan, S., Tsai,S., and Israel, M. A. Cytogenetic characterization of selected small roundcell tumors of childhood. Cancer Genet. Cytogenet., 21: 185-208, 1986.

47. Thiele, C, McKeon, C., Triche, T. J., Ross, R. A.. Reynolds, P., and Israël,M. Differential proto-oncogene expression characterizes histopathologicallyindistinguishable tumors of the peripheral nervous system. J. Clin. Invest.,«0:804-811, 1987.

48. Patel, K., Moore. S. E., Dickson, G., Roussel, R. J., Beverly, P. C., Kemshead,J. T., and Walsh, F. S. Neural cell adhesion molecule (NCAM) is the antigenrecognized by monoclonal antibodies of similar specificity in small-cell lungcarcinoma and neuroblastoma. Int. J. Cancer, 44: 573-578, 1989.

49. Cheung, N. K. V., Saarinen, V. M., Neely, J. E., Landmeier, B., Donovan,D., and Coccia, P. F. Monoclonal antibodies to a glycolipid antigen onhuman neuroblastoma cells. Cancer Res., 45: 2642-2649, 1985.

50. Saito, N., Yu, R. K., and Cheung, N. K. V. Ganglioside GD2 specificity ofmonoclonal antibodies to human neuroblastoma cell. Biochem. Biophys.Res. Commun., 127: 1-7, 1985.

51. Lashford, I. S.. Jones, D. H., Pritchard, J., Gordon, I.. Bretnach, F., andKemshead, J. T. Therapeutic application of radiolabelled monoclonal antibody UJ13A in children with disseminated neuroblastoma. Nati. Cancer Inst.Monogr., 3: 53-59. 1987.

52. Cheung, N. K. V., Landmeier, B., Neely, J., Nelson, A. D., Abramowsky, C,Ellery, S., Adams, R. B., and Miraldi, F. Complete tumor ablation withiodine 131-radioiabelled disialo-ganglioside GD2-specific monoclonal antibody against human neuroblastoma xenografted in nude mice. J. Nati. CancerInst., 77:739-745, 1986.

53. Cheung, N. K. V., Lazarus, H., Miraldi, F. D., Abramowsky, C. R., Kallick,S., Saarinen, U. M., Spitzer, T., Strandjord, S. E., Coccia, P. F., and Berger,N. A. Ganglioside GD2 specific monoclonal antibody 3F8: a phase I studyin patients with neuroblastoma and malignant melanoma. J. Clin. Oncol., 5:1430-1440,1987.

54. Roth, J., Zuber, C., Wagner. P.. Taatjes, D. J., Weisgerber. C., Heitz, P. U.,Goridis, C., and Bitter-Suermann, D. Reexpression of polysialic acid unitsof the neural cell adhesion molecule in Wilms tumor. Proc. Nati. Acad. Sci.USA, 85: 2999-3003. 1988.

55. Moore, S. E., and Walsh, F. S. Specific regulation of N-CAM/D2-CAM celladhesion molecule during skeletal muscle development. EMBO J.. 4: 623-630, 1985.

56. Walsh, F. S., and Moore, S. E. Expression of cell adhesion molecule. N-CAM, in diseases of adult human skeletal muscle. Neurosci. Lett., 59: 73-78, 1985.

1307

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

1991;51:1300-1307. Cancer Res   Gunhild Mechtersheimer, Martina Staudter and Peter Möller  Their Tumorsand CD57 in Human Neural and Striated Muscle Cells and in Expression of the Natural Killer Cell-associated Antigens CD56

  Updated version

  http://cancerres.aacrjournals.org/content/51/4/1300

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/51/4/1300To request permission to re-use all or part of this article, use this link

Research. on October 3, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from